OUP user menu

Ecology and Social Organization of a Tropical Deer (Cervus Eldi Thamin)

Myint Aung, William J. McShea, Sein Htung, Aung Than, Tin Mya Soe, Steven Monfort, Chris Wemmer
DOI: http://dx.doi.org/10.1644/1545-1542(2001)082<0836:EASOOA>2.0.CO;2 836-847 First published online: 17 August 2001


From 1995 to 1999, we conducted an ecological study of thamin (Cervus eldi thamin) at Chatthin Wildlife Sanctuary in central Myanmar; we maintained records on deer sightings and radiotracked 11 adult male and 8 adult female deer. Based on 747 sightings, a 0.63:1.0 adult male : female ratio and 0.51:1.0 fawn : adult female ratio were observed. Mean group size was variable (1.0–5.9 deer) and showed seasonal differences, with few groups observed in August–September and groups of ≤70 individuals observed in March–April. Based on the fixed-kernel method, annual home range was 9.04 km2 ± 5.67 SD and 7.25 km2 ± 3.45 SD for males and females, respectively. Thamin increased their seasonal home range during the hot–dry season, possibly in response to decreased forage and water availability and increased mating activity. The observed synchrony of estrous onset (March–April) and fawning (October–November) in female thamin is unusual for a tropical cervid species, but reproductive seasonality appears timed to balance fawn survival with doe nutrition in a monsoon environment.

Key words
  • Cervidae
  • Cervus eldi
  • fire ecology
  • predation
  • reproduction
  • seasonality
  • thamin

Life-history patterns of ungulates are influenced strongly by the annual cycle of energy needs (Bronson 1989) and predation (Geist 1998; Leslie et al. 1999). Social organization and mother–young interactions particularly are responsive to predation pressure in ungulate species in open habitats (Bowyer et al. 1998; Hirth 1977; Jarman and Jarman 1979; Kie 1999; Kie and Bowyer 1999). For species subjected to seasonal changes in vegetative cover, social group size and the extent of daily and seasonal movement are correlated negatively with the amount of vegetative cover (Geist 1998; Hirth 1977; Jarman 1974; Molvar and Bowyer 1994). Whereas vegetative cover for temperate deer species typically is reduced when ambient temperatures are low (i.e., winter), tropical deer species often inhabit ecosystems in which vegetation is burned seasonally during the hot season (Dhungel and O'Gara 1991; Dinerstein 1987; Mishra and Wemmer 1987; Schaller 1967).

The thamin (Cervus eldi thamin) is a tropical cervid that has been extirpated throughout most of its historic range (McShea et al. 1999; Wemmer 1998), but it occurs still in the central plains of Myanmar. It is a relatively large cervid (70–130 kg) that once inhabited open, dry woodlands throughout Southeast Asia (Salter and Sayer 1986). Historic range of the thamin was shared with 3 large predators: tiger (Panthera tigris), leopard (Panthera pardus), and dhole (Cuon alpinus), although only the latter 2 still are abundant in the region. Our purpose was to study the natural history (i.e., reproduction, survival, social organization, movements) of this tropical cervid in light of current knowledge for captive populations and other tropical cervids.

In contrast to the well-established relationship between photoperiod and seasonal reproduction in temperate species, ecological factors responsible for modulating reproductive seasonality in many tropical deer species are poorly understood (Geist 1998). Descriptions of reproduction and behavior of thamin have relied on anecdotal observations of wild populations (Evans 1894; Salter and Sayer 1986; Yin 1967) and detailed research on captive animals (Monfort et al. 1990, 1993; Wemmer and Grodinsky 1988; Wemmer and Montali 1988). For example, behavioral and physiological studies of captive female thamin have shown them to be seasonally polyestrous with onset of the estrous cycle in late winter or early spring in North America; 80% of births occur between September and November (Monfort et al. 1993; Wemmer and Grodinsky 1988).

Captive thamin have been maintained in temperate zones for >50 years, and although maladapted to northern temperate latitudes (as far north as 48°N), they continue to exhibit seasonal reproductive rhythms (peak breeding March–April, peak births October–November) identical to patterns of their subtropical counterparts (Monfort et al. 1990). Thamin may have developed endogenous reproductive rhythms that have evolved in response to seasonal fire and rainfall—dominant environmental events in the dry forests of Southeast Asia. Although several studies (Blouch 1987; Branan and Marchinton 1987; Smith et al. 1996) have demonstrated a relationship between annual rainfall patterns and timing of the fawning season of subtropical and tropical deer species, we can infer from the lack of interspecies uniformity that adaptations to environmental conditions are based on a variety of regulatory mechanisms. No attempt has been made to examine the reproductive pattern of thamin relative to the ecology of their native habitat.

Based on limited field data, wild thamin typically are sighted alone, except during the spring months, when deer have been observed in large groups of both sexes (Salter and Sayer 1986). A population survey conducted in central Myanmar reported group sizes of 1–20 animals, an adult male : female ratio of 1.00:1.59, and a doe : fawn ratio of 1.00:0.54 (Salter and Sayer 1986). Although formally listed as endangered (International Union for the Conservation of Nature 1997), populations of thamin still can be found within some portions of the dry central plain of Myanmar (McShea et al. 1999), but ecological information necessary for conservation is lacking.

Our study examined the interrelationships between life-history patterns and annual fire- and weather-induced changes in a dry dipterocarp forest ecosystem of central Myanmar. These data will improve our understanding of how tropical cervid species adapt to seasonal change in tropical deciduous forests and aid in development of comprehensive management plans to conserve thamin.

Materials and Methods

Study area

Chatthin Wildlife Sanctuary (CWS) lies on the northern edge of the central plains of Myanmar (95°24′–95°40′E, 23°30′–23°42′N) and is situated in Kanbalu and Kawlin townships, Sagaing Division, Myanmar. The sanctuary encompassed 310 km2, but most observations occurred in the eastern part of the sanctuary and adjoining agricultural areas (204 km2). Regions of South Asia that experience monsoon rains are characterized by 3 seasons (Dhungel and O'Gara 1991): cool–dry (October–January), hot–dry (February–May), and rainy or monsoon (June–September). Weather records from CWS indicated that average precipitation was 35 cm over the 4 years of the study (February 1995–January 1999; Fig. 1). High temperatures (35–45°C) during the hot–dry season persisted through the middle of the monsoon but moderated (<30°C) in December–January. A series of north–south and east–west transects were measured and marked along the eastern half of the sanctuary where the study was conducted. Those transects created a grid of 91 1.5- by 1.5-km blocks that were used to reference deer sightings.

Fig. 1.

Pattern of rainfall (histograms) and maximum (solid circle) and minimum temperatures (open circle) during the study at Chatthin Wildlife Sanctuary in central Myanmar

The predominant habitat was a deciduous dipterocarp or monsoon forest known as indaing (Stamp 1925), which was dominated by Dipterocarpus tuberculatus and its associates Pentaceme siamensis, Shorea oblongifolia, Cycas siamensis, and Dillenia parviflora. Denser mixed-deciduous forests were found along streams and at higher elevations. Grassy, seasonally flooded plains (lwins) were interspersed through the forest, and CWS was bordered by agricultural land, with rice as the dominant crop. The lwins supported a variety of grasses, including dense stands of Saccharum and Imperata with a height of >2 m.

Sighting records

All thamin sightings were recorded, including total number of deer, sex, age, habitat, location, date, and time of sighting. Deer were classified as adult male, adult female, juvenile, or unknown. Fawns were recognized by their small size and spotted pelage. During their antlerless stage (rainy season, June–September), males could not be distinguished readily from females, and sightings were confounded by dense understory vegetation during the monsoon. Thus, no sex was recorded for deer sighted during the monsoon. For all other seasons, males were classified as being in velvet or hard antler.

Deer capture

Thamin deer were captured from January 1995 through August 1998. We used 400 m of drive nets (in 30-m sections) made from nylon with a 15-cm mesh. The nets were set as V-shaped funnels, and a double beating system was used. Captured deer were ear-tagged and weighed, and several body measurements were taken, including total length, shoulder height, chest girth, and antler length. Fawns and yearling deer were released after measuring, whereas each adult male and female thamin were fixed with a radiocollar.

Observations of radiocollared deer

Nineteen adult thamin (11 males, 8 females) were fitted with radiocollars. Each radiocollar had a transmitter (300 g; Model 500 Telonics, Inc., Tucson, Arizona), with a range of about 2 km on the ground (3–4 km from an elevated position) and an average battery life of >2 years. Radiocollared deer were located 4 days/week during daylight hours (0700 h and 1900 h) using a 4-element Yagi antenna and a 12-channel radioreceiver (Telonics, Inc.). Tracking was conducted on foot, and deer were approached to within 100 m, circled, and located. For each deer located, date, time, block number, habitat type, and dominant flora species were recorded. For deer that could be seen, behaviors (e.g., feeding, resting), antler stage (velvet, hard or dropped), group size and composition, and plant species that thamin were eating also were recorded. When deer fled before visual sighting, locations were recorded only if deer were estimated to have been ≤100 m of the observer before it fled. Location of deer within each block was placed on a 1:50,000 map of the sanctuary using landmarks, trails, and marked transects as references.

Statistical analysis

Data were analyzed using SAS (version 6.2; SAS Institute Inc. 1987) and ArcView (version 3.1; Environmental Systems Research Institute, Inc., Redlands, California). All locations of radiocollared deer were overlaid on a map of CWS that was georeferenced with UTM coordinates of significant landmarks obtained with a Trimble ProXR global-positioning unit (Trimble Navigation, Sunnyvale, California). The fixed-kernel method with least squares cross validation (Worton 1986, 1989) was used to determine a 95% polygon for home ranges (spatial analyst extension for ArcView software). Least-squares cross validation was selected for the smoothing parameter because it produced the smallest bias in estimates of area that use the fixed-kernel method (Seaman and Powell 1996). We also calculated home-range size using a convex polygon encompassing the outermost animal locations to compare with previous studies. Home-range calculations were based on all radiolocations of individual deer in each of the 3 seasons and over an entire year. The minimum samples necessary to compute home ranges were estimated using linear regression (SAS Institute Inc. 1987), and no significant relationship existed between estimated home-range size and number of locations when number of locations used was >45 (F = 0.01, d.f. = 1, 63, P > 0.1). Number of locations used to determine home ranges ranged from 49 to 88 for seasonal estimates and 143 to 235 for yearly estimates. Centers of activity were calculated weekly using the 25% polygon of the fixed-kernel method to quantify movement. Analysis of variance was used to compare home-range sizes and movement between sexes and seasons. The standard error (SE) is given for each mean unless indicated otherwise.


Home range

Average sizes of the annual home ranges (fixed kernel) were 9.04 ± 5.67 km2 for males and 7.25 ± 3.45 km2 for females (Table 1) and varied greatly among individuals of the same sex (males, 1.85–18.37 km2; females, 2.6–11.11 km2). There were no differences between sexes in annual home ranges (partial F = 0.69, d.f. = 1, 45, P > 0.1), but home ranges differed seasonally (partial F = 4.80, d.f. = 2, 44, P = 0.013), being larger in the hot–dry season (Table 1). There was no interaction between sex and season (partial F = 0.91, d.f. = 2, 44, P > 0.1). Males and females exhibited site fidelity for seasonal home ranges, showing a high degree of overlap (Fig. 2).

View this table:
Table 1.
Fig. 2.

Representative home ranges (95% fixed-kernel estimate with least-squares cross validation) for A) female and B) male thamin throughout 1 year. Seasons are shaded the same for each deer: the hot–dry season is clear, rainy season is speckled, and cool season is diagonal hatching. Seasonal home ranges are overlayed for each animal, and the smaller home range is on top. Animals were selected to minimize overlap between individuals; arrows connect the ranges of animals with disjunct distributions

Annual home-range estimates based on convex polygons did not differ between male (16.76 ± 3.11 km2) and females (12.47 ± 1.97 km2; F = 1.97, d.f. = 1, 14, P = 0.28). That may have been due to small samples and large individual variation for males (8.42–31.71 km2; n = 9) and females (3.25–19.33 km2; n = 7).


Center of activity for each radiocollared animal shifted an average of 450 m/week, with a maximum shift of >900 m during the hot–dry season (Table 2). The magnitude of those shifts differed among seasons (partial F = 15.66, d.f. = 2, 807, P = 0.0001), but sexes did not differ (partial F = 2.21, d.f. = 1, 808, P = 0.14). There was an interaction between season and sex (partial F = 3.69, d.f. = 2, 807, P = 0.02); females moved farther than males during the cool–dry season.

View this table:
Table 2.

Some movements appeared to be regular temporally. For example, male 210 moved >9 km with the onset of each hot–dry season from 1996 through 1998. For that male, timing of movement was consistent; departure from the cool–dry season range occurred in February (mean departure date, 19 February ± 8 days), and departure from the hot–dry season range occurred in early May (mean departure date, 6 May ± 4 days). No other animals showed such large movements, but 8 individuals (4 males and 4 females) showed seasonal shifts in home range.

Group size

Male thamin spent most of the year alone, and females associated only with their fawns (Table 3). Of 747 sightings when sex could be identified, 29% (n = 277) were of a single adult, and 15% (n = 143) were of a single adult female with fawn. The pattern changed dramatically during the hot–dry season, when groups of ≤70 individuals were observed. Mean group size peaked in April (5.9 ± 8.3) and decreased slowly to 1.1 ± 0.3 through September (Table 3). Males were 1st observed in groups in November, and multimale groups were found mostly during the late cool–dry and early hot–dry season (December–March).

View this table:
Table 3.

We made 2,760 observations of 19 radiocollared deer (838 hot–dry, 806 rainy, 1,116 cool–dry). Mean group size was highest in the hot–dry season and lowest in the rainy season (Fig. 3). Male radiocollared deer usually were solitary, but the proportion of observations of solitary deer varied seasonally: 62% (n = 444), hot–dry season; 90% (n = 490), rainy season; 74% (n = 777), cool–dry season. As with sightings, radiocollared males were observed with female groups from the late cool–dry season through the hot–dry season (Fig. 3).

Fig. 3.

Mean group size (histograms) and percentage of single deer sightings (open circle) from 2,760 sightings of 19 radiocollared thamin and their associates. Single deer observations exclude nonadult sightings. Presence of antlers (solid circle) is based on monthly observations of 11 radiocollared males

Radiocollared females associated with their fawns throughout the year. Females were found alone or with their fawn 55% (n = 394), 84% (n = 316), and 80% (n = 339) of the time during the hot–dry, rainy, and cool–dry seasons, respectively.

Sex ratio and age composition

The proportions of males, females, and fawns in the population were calculated from sightings of 2,035 individuals in 747 groups. We were unable to determine the sex ratio during the rainy season (June–September) because of lower visibility and the antlerless condition of males. Exclusive of the rainy season, the adult sex ratio (male : female) for this population was 0.63:1.00, with no consistent differences between the hot–dry and cool–dry seasons (Table 3).

The fawn : adult female ratio was highest during February (0.71:1.00) and declined to a low in October (Table 3). We assumed that the ratio in February more closely approximated productivity of this population because the rate of mortality was probably higher in fawns than adult females, and reproduction was seasonal.


The annual reproductive life cycle was inferred from direct observations of antler development in males, group composition, and timing of fawning. For deer with synchronous mating seasons, male–male interactions typically peak before male–female interactions (Lincoln 1971). The greatest number of all male groups was observed from November to January, just before the observed peak in mixed-sex group composition (February–May; Table 3). Whereas overt mating behaviors (i.e., mounting and copulation) were not observed in radiocollared males, those males tended females in March–April (n = 12).

Visual observations of 11 radiocollared males were used to track the annual antler cycle. All males had hard antlers from late December to May; 75% possessed hard antlers during early June, 58% in late June, 14% in early July, and none by late July (Fig. 3). Thus, antlers were shed between the last week of May and mid-July. Antlers were observed in the velvet stage in July (21%) and August (86%). All males were seen with velvet in September–October, after which observations of deer with velvet declined steadily; 88% of male thamin were in hard antler by early December.

Six of 8 radiocollared females produced fawns. Of the other 2 females, 1 was with a fawn, and the other was with a yearling male, when first captured. Fawns were observed first in December, but fawns never were observed at their bedding sites. Of those females that were radiotracked for >1 birthing season (December), only 1 produced a 2nd fawn, and that was after an interval of 2 years. Although single adult females were sighted with 2 fawns, no twins were observed in the sample of radiocollared animals.


We tracked the 19 thamin for 279 animal-months during the 48-month study. During the 4-year study, 4 radiocollared deer were killed by dholes, 1 was killed by a hunter, 1 died of unknown causes, 2 disappeared, and 2 dispersed out of the study area (a female and yearling male captured together). Most deaths (4 of 6), all dispersals, and all disappearances occurred during the cool–dry season.


Social organization and life-history traits of thamin differed from other tropical cervids in the extent to which they exhibited large daily and seasonal movements, tightly synchronized seasonal breeding, seasonal variation in group size, and low reproductive output by females. Our estimate of home-range size for thamin was significantly greater than other South Asian cervids (Table 4). Estimates of the home ranges (minimum convex polygon) of hog deer (Axis porsinus) in Nepal were 0.8 km2 for males and 0.6 km2 for females (Dhungel and O'Gara 1991). Similar estimates for chital (Axis axis) in Nepal were 3.0 km2 in male and 2.1 km2 in female (Mishra and Wemmer 1987). Most of these range differences with thamin are probably due to differences in body size (Geist 1998). The barasingha (Cervus duvauceli) and chital are closest in body size to thamin, but their home ranges are markedly smaller (Table 4). Both species form year-round social herds and concentrate near water (Schaller 1967); both behaviors may reduce their movements more than thamin. Lack of sex-related differences in size of home range of thamin is similar to hog deer (Dhungel and O'Gara 1991) but is exceptional compared with most tropical cervids (Geist 1998; Mishra 1982). Individuals of both sexes demonstrated site fidelity (Fig. 2), and home-range estimates did not increase beyond 45 positions, which took about 2 months to obtain for each radiocollared thamin.

View this table:
Table 4.

In many tropical ungulate species, seasonal changes in the environment determine movement patterns and home-range size (Dinerstein 1987; Geist 1998; Jarman and Jarman 1979). This deer exhibits large movements because the center of activity for each deer showed average shifts of up to 1 km. Thamin forage primarily on the new grass that follows localized fires, low-density forbs, and widely dispersed fruits (McShea et al. 2001)—all resources that are probably quickly exploited within any single patch.

In both sexes of thamin, movements and home-range size during the hot–dry season were significantly larger than during the other seasons. Three possible explanations are reduced availability of water and forage crop, increased activity by villagers living in and around the sanctuary, and the mating season.

Reduced or patchy availabilities of forage and water during the hot–dry season may induce an increase in home-range size and movements (cf. Kie 1999). For thamin, water availability is quite limited during the hot–dry season, when a only a few widely dispersed permanent water holes are present CWS (McShea et al. 2001). For agricultural land surrounding CWS, crops are planted when the monsoon begins (June–July), and most crops, a preferred food resource of thamin (McShea et al. 2001), are harvested after October during the cool–dry season. We speculate that the increase in home range of both sexes reflects the necessity to broaden their search for these limited food and water resources.

Human activities may have short-term or seasonal impacts on deer movements. We observed dramatic increases in movements in some in radiocollared deer (>2,300 m/week) during the rainy and cool–dry seasons when villagers collected forest products (e.g., mushrooms, thatch-grass) from CWS. Domestic dogs often accompany villagers on these collecting trips, and dogs are used by villagers to hunt thamin. During the cool–dry season, when female thamin spent more time near human developments (i.e., near croplands), their average movements exceeded that of males. Additionally, fewer deer were found near the sanctuary boundary (near human developments) during the hot–dry season, a time when vegetation cover was scarce and villagers, freed from cultivating their fields, were more active around the sanctuary boundaries.

Thamin are seasonal breeders, and peak mating occurred during the hot–dry season. Increased movements during rut are common for ungulates (Holzenbein and Schwede 1989; Relyea and Demarais 1994). The concentration of forage in emergent grass shoots following annual fires may cause increased movement because animals follow the erratic patterns of the fires (Dinerstein 1987). Fire-induced growth of grasses coincides with the mating season of thamin. Whereas male thamin exhibit an obligate “catabolic” decline in body weight and voluntary food intake during rut (Monfort et al. 1993), females probably continue feeding activity and searching for ephemeral food (and water) sources. Thus, the increased group sizes of thamin during the mating season likely represented feeding aggregations of female deer that were exploiting newly sprouted patches of grass and male deer who are following the estrous females. Hog deer also form large groups following the burning of their grassy habitat (Dhungel and O'Gara 1991). Chital (Mishra 1982) and barasingha (Schaller 1967) have been reported in large groups, but their group size does not vary seasonally.

An alternate, but less likely, explanation for the large aggregations of thamin is they may be using each other as “cover” in the aftermath of extensive fires that destroy natural vegetative cover (Barrette 1991). Several ungulates are known to increase group size with increased distance from cover (Odocoileus virginianusHirth 1977; Alces alcesMolvar and Bowyer 1994; Ovis dalliRachlow and Bowyer 1998), and this may be an antipredator defense strategy (Kie 1999). Dhole caused significant mortality within this population, so predator defense may be important.

Males no doubt exploit female aggregations to maximize their reproductive fitness. We have experimentally demonstrated that male presence (either direct or indirect exposure) augments ovarian function (Hosack et al. 1998) and synchronizes ovulation onset (Hosack et al. 1999) in captive female thamin; in fact, ovulation was delayed in male-isolated females. Male chemosignals may act in concert with the “tending-bond” mating strategy used by thamin males, whereby they maintain transient but close proximity to females and defend them from other males until mating occurs (Monfort et al. 1993). Mixed-sex aggregations of thamin also would increase the likelihood of females being exposed to secretions of the subcaudal scent gland and latrines in which males repeatedly deposit fecal pellets and urine (Wemmer and Montali 1988). The prolonged proximity of males and females may cause females to ovulate, thereby providing a mechanism to shorten the breeding season. Fawn mortality and female fecundity can be decreased in red deer (Cervus elaphus) by 1% for each day conception is delayed compared with the preceding year (Clutton-Brock et al. 1983). Large aggregations of thamin may, therefore, ensure early conceptions and maximize chances that births will occur when environmental conditions are optimal for offspring survival.

The exact timing of thamin births was difficult to determine in the field because mothers hide their young immediately after birth, like other cervid species (Blakeslee et al. 1979; Lent 1974). However, the ratio of fawns per female increased after October, and radiocollared females first were observed with fawns in December. If fawns were born in November–December and gestation lasted about 34 weeks (Monfort et al. 1993), this would indicate that mating occurred in March–April. Although systematic field data still are lacking for most tropical cervids, the restricted mating season for thamin appears to be comparatively short (Table 4). A 60–90-day birthing season for thamin would contradict Rutberg's (1987) prediction that low-latitude cervids that use a “hider” strategy to protect neonates will exhibit extended birthing seasons (predicted birthing season of thamin using Rutberg's regression models is 108–209 days).

Whereas our knowledge about how seasonal reproduction is modulated in temperate cervids is strong (Bronson 1989; Geist 1998), comparatively little is known about reproductive seasonality in tropical cervids. For all cervids, timing of birth is a trade-off between nutrition and predation (Bowyer et al. 1998, 1999). Because mid- to late lactation is the most energetically costly period for any female, we would predict that thamin should have evolved adaptive mechanisms for matching peak energetic requirements to seasonal fluctuations in forage quality and fawns should be weaned at a time of maximal forage quality (Bronson 1989).

Contrary to these predictions, thamin wean their young during the hot–dry season, when forage quality is at its lowest. Thus, we speculate that parturition of thamin has evolved as a trade-off between fawn safety and energetic needs of the dam and fawn (Bowyer et al. 1999). Birthing during the cool–dry season, when flood waters have receded (after the monsoon) and dense ground vegetation provides adequate shelter for fawns, ensures that weaned fawns (4–5 months old at the beginning of the hot–dry season) will have sufficient mobility to reduce vulnerability to predation and exploit shifting food and water resources (Fig. 4). In contrast, fawn survival during the hot–dry season would be compromised by widespread fires, increased susceptibility to predation due to lack of vegetative cover, and poor female nutrition (i.e., lactational stress). Similarly, extensive flooding (>50% of CWS is inundated with water for extended periods) during the monsoon would increase risk of neonatal mortality due to fawn drowning, and wet conditions could impair thermoregulatory function in neonates. Births in other cervids also are timed to coincide with the end of the rainy season, when conditions for fawn survival are considered favorable.

Fig. 4.

Schematic representation of reproductive life cycle of the thamin at Chatthin Wildlife Sanctuary in central Myanmar

There was ample evidence of predators of thamin within CWS. Although leopards were reported by local villagers, the only predators observed by CWS staff were dholes, jackals, and occasionally feral dogs. There were 9 recorded instances of predation of thamin by dholes (total from radiocollaring and sighting data), and 6 of those deaths occurred during the cool–dry season. Thamin evolved in Southeast Asia in the presence of people and are still hunted illegally by villagers, primarily with firearms and spears. Estimates of 2–3 killings/day have been reported (Milton and Estes 1963), and Salter and Sayer (1986) counted 21 deer killed in 1 day at CWS. Whereas poaching certainly still occurs, more information is needed to quantify the extent of poaching activities. Poachers that shot 3 thamin in 1 night were apprehended by CWS staff, and 1 radiocollared deer was killed by a poacher during the study. Conflict between needs of humans and animals for limited resources remains the greatest threat to survival of thamin, but human disturbance should be considered along with monsoonal flooding, fires, and predation as major factors shaping life-history traits of thamin.


This work was supported by funds from the National Geographic Society, nonfederal trust funds of the Smithsonian Institution, Friends of the National Zoo, and the Wildlife Conservation Society. British Airways supported travel through its Giving Conservation a Lift Program. Field assistance was provided by many individuals, but in particular by T. T. Yu and M. Win. Permits and clearances were arranged by U Uga and U Than Nwai. K. Koy, J. Müller, and D. Poszig provided computer assistance. We thank the Nature and Wildlife Conservation Division of the Department of Forests, Myanmar, for permission to conduct this work.

Literature Cited

View Abstract